Unusual Reactions of the Model Carcinogen N-Acetoxy-N-acetyl-2-amino-α-carboline

Article abstract of DOI:10.1021/jo034505u

The aqueous solution reactions of the title compound, 1, were examined for comparison to those previously reported for another model carcinogen N-pivaloyloxy-2-amino-α-carboline, 2. Both of these are models for the ultimate carcinogenic metabolites of 2-amino-α-carboline (AαC), a food-derived heterocyclic amine mutagen and carcinogen. The present study was undertaken to determine the effect of the N-acetyl group on the chemistry of such compounds. The N-acetyl group slows down N-O bond cleavage by a factor of (5.5 × 10³)-fold. This allows other reactions not observed in 2, or in other model carcinogens, to be observed. Among these are acyl-transfer reactions to the aqueous solvent, both uncatalyzed and catalyzed by N ₃^-. In addition, the conjugate acid of 1, 1H⁺, is subject to a spontaneous decomposition not previously observed in other esters of heterocyclic hydroxylamines or hydroxamic acids. This reaction yields the hydroxylamine, 5, and does so without the intermediacy of the hydroxamic acid, 3, and with ¹⁸O exchange from the solvent into the hydroxylamine O. This unique reaction may be caused by an intramolecular proton donation by the pyridyl N-H to the amide carboxyl that catalyzes an intramolecular nucleophilic attack by the carboxyl O of 1H⁺. A nitrenium ion pathway can still be detected for 1, but, unlike 2 and related esters, this reaction is in competition with other processes throughout the pH range of the study.

Full text of DOI:10.1021/jo034505u

VOLUME 68, NUMBER 26
DECEMBER 26, 2003
© Copyright 2003 by the American Chemical Society
Un u su a l Rea ction s of th e Mod el Ca r cin ogen
N-Acetoxy-N-a cetyl-2-a m in o-r-ca r bolin e
Michael Novak* and Thach-Mien Nguyen
Department of Chemistry and Biochemistry, Miami University, Oxford, Ohio 45056
novakm@muohio.edu
Received April 18, 2003
The aqueous solution reactions of the title compound, 1, were examined for comparison to those
previously reported for another model carcinogen N-pivaloyloxy-2-amino-R-carboline, 2. Both of
these are models for the ultimate carcinogenic metabolites of 2-amino-R-carboline (ARC), a food-
derived heterocyclic amine mutagen and carcinogen. The present study was undertaken to determine
the effect of the N-acetyl group on the chemistry of such compounds. The N-acetyl group slows
down N-O bond cleavage by a factor of (5.5 × 103)-fold. This allows other reactions not observed
in 2, or in other model carcinogens, to be observed. Among these are acyl-transfer reactions to the
aqueous solvent, both uncatalyzed and catalyzed by N3-. In addition, the conjugate acid of 1, 1H+,
is subject to a spontaneous decomposition not previously observed in other esters of heterocyclic
hydroxylamines or hydroxamic acids. This reaction yields the hydroxylamine, 5, and does so without
the intermediacy of the hydroxamic acid, 3, and with 18O exchange from the solvent into the
hydroxylamine O. This unique reaction may be caused by an intramolecular proton donation by
the pyridyl N-H to the amide carboxyl that catalyzes an intramolecular nucleophilic attack by the
carboxyl O of 1H+. A nitrenium ion pathway can still be detected for 1, but, unlike 2 and related
esters, this reaction is in competition with other processes throughout the pH range of the study.
In tr od u ction
We have been involved in a long-term study of the
aqueous solution chemistry of ester derivatives of het-
erocyclic N-arylhydroxylamines.1-3 These are the puta-
tive metabolites of heterocyclic amines generated during
the cooking of protein-containing foods.4,5 Our studies
* Corresponding author.
(1) Novak, M.; Xu, L.; Wolf, R. A. J . Am. Chem. Soc. 1998, 120,
1643-1644.
(2) Novak, M.: Kazerani, S. J . Am. Chem. Soc. 2000, 122, 3606-
3616.
(3) Novak, M.; Toth, K.; Rajagopal, S.; Brooks, M.; Hott, L. L.;
Moslener, M. J . Am. Chem. Soc. 2002, 124, 7972-7981.
(4) Eisenbrand, G.; Tang, W. Toxicology 1993, 84, 1-82. Hatch, F.
T.; Knize, M. G.; Felton, J . S. Environ. Mol. Mutagen. 1991, 17, 4-19.
Layton, D. W.; Bogen, K. T.; Knize, M. G.; Hatch, F. T.; J ohnson, V.
M.; Felton, J . S. Carcinogenesis 1995, 16, 39-52. Sugimura, T. Mutat.
Res. 1985, 150, 33-41. Sugimura, T. Environ. Health Perspect. 1986,
67, 5-10. Ohgaki, H.; Takayama, S.; Sugimura, T. Mutat. Res. 1991,
259, 399-410. Felton, J . S.; Knize, M. G.; Shen, N. H.; Andersen, B.
D.; Bjeldanes, L. F.; Hatch, F. T. Environ. Health Perspect. 1986, 67,
17-24.
(5) Watanabe, M.; Ishidate, M.; Nohmi, T. Mutat. Res. 1990, 234,
337-348. Raza, H.; King, R. S.; Squires, R. B.; Guengerich, F. P.;
Miller, D. W.; Freeman, J . P.; Lang, N. P.; Kadlubar, F. F. Drug Metab.
Dispos. 1996, 24, 395-400. Pfau, W.; Brockstedt, U.; Schulze, C.;
Neurath, G.; Marquardt, H. Carcinogenesis 1996, 17, 2727-2732. Pfau,
W.; Schulze, C.; Shirai, T.; Hasegawa, R.; Brockstedt, U. Chem. Res.
Toxicol. 1997, 10, 1192-1197. King, R. S.; Teitel, C. H.; Kadlubar, F.
F. Carcinogenesis 2000, 21, 1347-1354. Frederiksen, H.; Frandsen,
H. Pharmacol. Toxicol. 2002, 90, 127-134.
have shown that these esters generate heterocyclic
-
N-arylnitrenium ions that react readily with N3 and
with 2′-deoxyguanosine.1-3 For reasons of synthetic ease
and stability of the model esters we have occasionally
utilized esters of heterocyclic N-acetyl-N-arylhydroxyl-
amines.3
Studies with carbocyclic analogues have shown that
heterolytic N-O bond cleavage is slowed by a factor of
up to 105 by replacement of NH with NAc in the esters,
but the fundamental nature of the reaction is unal-
tered.6,7 The chemistry of the resulting nitrenium ion is
only moderately altered by this replacement.6,7 Sites of
reaction with nucleophiles are not altered, and the
10.1021/jo034505u CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/26/2003
J . Org. Chem. 2003, 68, 9875-9881
9875

Novak and Nguyen
TABLE 1. Kin etic a n d Th er m od yn a m ic P r op er ties for 1 w ith Com p a r ison s to 2
pKa
a
a
compd
kA (s-1
)
kB (s-1
)
kinetic fita
titrationb
1c
2d
(7.8 ( 0.4) × 10-4
(9.8 ( 0.4) × 10-5
-0.08 ( 0.07
1.73 ( 0.06
-0.04 ( 0.09
(8.3 ( 0.2) × 10-2
compd
∆Hq (kcal/mol)e
16.4 ( 0.6
∆Sq (cal/K‚mol)e
-25 ( 2
kB(2)/kB(1)f at 20 °C
kBKa/kA (M) from eq 2g
1
(5.5 ( 0.8) × 103
0.12 ( 0.02
a
b
From a weighted least-squares fit to eq 1. Values reported with their estimated standard deviations obtained from the fit. From a
fit of initial absorbance at 305 nm vs pH to a standard titration equation. c Conditions: 5 vol % of CH3CN-H2O, µ ) 0.5, T ) 40 °C.
Conditions: 5 vol % of CH3CN-H2O, µ ) 0.5, T ) 20 °C. Source: ref 2. e From a fit of ln(kB/T) at pH 7.0 vs 1/T at 40, 50, and 60 °C.
d
f From kB for 1 extrapolated to 20 °C and kB for 2 reported in ref 2. Fit to eq 2 from product yield data for 5 (see Figure 3).
g
quantitative measure of reactivity and selectivity, kaz/ks,
the ratio of the second-order rate constant for reaction
-
of the ion with N3 and the first-order rate constant for
reaction with the aqueous solvent, is reduced by less than
an order of magnitude by the N-acetyl group.6,7
No comparative data have been available for hetero-
cyclic N-arylnitrenium ions. For this reason we initiated
a study of the chemistry of N-acetoxy-N-acetyl-2-amino-
R-carboline, 1, for comparison with our previously pub-
lished study of the chemistry of N-pivaloyloxy-2-amino-
R-carboline, 2.2 Both of these compounds are models for
mutagenic and carcinogenic metabolites of the heterocy-
clic amine 2-amino-R-carboline, ARC.5 Our results show
that the N-acetyl group can have a much more profound
effect on the chemistry of these heterocyclic esters than
anticipated from the results with carbocyclic esters.
F IGURE 1. Log k1 vs pH for 1 and 2. Rate constants for 1
come from HPLC (b) and UV (O) measurements at 40 °C. Data
for 2 (2) are from UV measurements at 20 °C reported in ref
Resu lts a n d Discu ssion
2. Data were fit to eq 1 for 1 and to the second term of eq 1 for
2.
Decomposition kinetics of 1 were monitored at 40 °C
in 5 vol % of CH3CN-H2O. In the pH range (pH >0.6)
ionic strength, µ, was maintained at 0.5 with NaClO4 and
plot of log k1 for 1 vs Ho or pH (Figure 1) is qualitatively
pH was maintained with HClO4 (pH <3.0) and with
and quantitatively quite different from the plot of log k1
HCO2H/HCO2-, AcOH/AcO-, and H2PO4-/HPO4 buff-
ers. In more acidic solutions (-2.0 e Ho e 0.2) no attempt
was made to maintain constant µ, and acidity was
2-
for 2 obtained at 20 °C from the previously reported
study.2 The data for k1 of 1 were fit well by eq 1,
maintained with HClO4. Ho values were obtained from
k1 ) (10(-x)kA + KakB)/(Ka + 10(-x)
)
(1)
the compilations of Paul and Long as revised by Yates
and Wai.8
where x ) Ho or pH. Equation 1 indicates that both 1
and 1H+, its conjugate acid, are subject to spontaneous
uncatalyzed decomposition. The data for 2 were previ-
ously fit to the second term of eq 1 (KakB/(Ka + 10(-x))), a
term consistent with rate-limiting N-O bond cleavage.2
Kinetic parameters obtained from the fits are reported
in Table 1. The kinetically obtained pKa for 1H+ is in
excellent agreement with the pKa obtained by fitting
initial absorbance of 1 at 305 nm vs pH (or Ho) to a
standard titration equation (Table 1). The pKa of 1H+ is
about 1.8 units more negative than that for 2H+ (Table
1), while kB for 2 is (5.5 × 103)-fold larger than kB for 1
extrapolated to 20 °C from data taken at 40, 50, and 60
°C in pH 7.0 phosphate buffers. The effect of the N-acetyl
group on pKa is consistent with the electron-withdrawing
effect of the acetyl group relative to H. The rate effect
on kB is also consistent with expectations for a rate-
limiting transition state that involves generation of a
positive charge on the exocyclic N.6,7 The sign and
magnitude of ∆Sq (Table 1) for kB suggest a highly
ordered transition state that is consistent with restricted
rotation about the C-N bond of an incipient nitrenium
Kinetics monitored by UV methods generally did not
fit well to a simple first-order rate equation, but could
be fit to consecutive first-order rate equations containing
two or three rate constants. Kinetics monitored by HPLC
(Table S-1 of the Supporting Information) showed that 1
decomposed in a first-order fashion throughout the Ho
and pH range of the study, but that two initial reaction
products, the hydroxamic acid, 3, and the amide, 4,
underwent further hydrolysis into the corresponding
hydroxylamine, 5, and the amine, 6, respectively (Tables
S-2 and S-3 of the Supporting Information).
Table S-1 shows that the rate constant k1 for the
decomposition of 1 obtained from HPLC measurements
is in good agreement with a rate constant, also assigned
as k1, observed by UV methods. The data show that k1 is
buffer independent, but does exhibit pH dependence. A
(6) Novak, M.; Kahley, M. J .; Eiger, E.; Helmick, J . S.; Peters, H.
E. J . Am. Chem. Soc. 1993, 115, 9453-9460.
(7) Novak, M.; Rajagopal, S. Adv. Phys. Org. Chem. 2001, 36, 167-
254.
(8) Paul, M. A.; Long, F. A. Chem. Rev. 1957, 57, 1-45. Yates, K.;
Wai, H. J . Am. Chem. Soc. 1964, 86, 5408-5413.
9876 J . Org. Chem., Vol. 68, No. 26, 2003

N-Acetoxy-N-acetyl-2-amino-R-carboline
F IGURE 2. Plot of the time course for appearance of 5 (b)
and 3 (insert, 2) during the decomposition of 1 at pH 1.10:
(solid lines) the fits of the data to a consecutive first-order rate
equation containing two rate constants; (- -) the component
of the time course of 5 attributable to hydrolysis of 3; (- - -)
the component of the time course of 5 attributable to the fast
process not involving 3; (- - - -) the predicted time course for
the appearance of 5 if all of it were produced by hydrolysis of
3 (see ref 10).
F IGURE 3. Plot of product yields vs pH for 3, 4, 5, and 6.
Symbols are identified in the figure. The line is a least-squares
fit of the data for 5 to eq 2.
k1 measured for the disappearance of 1. Only 30% of 5 is
generated in a slower process consistent with the mag-
nitude of k2, the rate constant for hydrolysis of 3 (Tables
S-2 and S-3). Figure 2 shows that the time course for the
appearance of 5 would be very different if all of it were
generated from hydrolysis of 3.10 A similar observation
was made at pH 1.92 except that in this case 51% of 5 is
formed without the intermediacy of 3. At higher pH, 3 is
stable enough (Table S-3) that no significant decomposi-
tion of it occurs (<5%) during the time it takes for the
decompositon of 1 to reach 10 half-lives. The small
amount of 5 observed in this pH range is formed with a
rate constant consistent with k1, not by slow decomposi-
tion of 3. At pH <1.0, 5 is the major reaction product
(Figure 3). In this pH range it is generated via a first-
order process with a rate constant equivalent to k1 (Table
S-2). Although 3 decomposes fairly rapidly in this pH
range (Table S-3), it would have been detected by HPLC
if generated in significant quantities during the decom-
position of 1. It appears that in this pH range all of the
observed 5 is generated directly from 1 without the
intermediacy of 3. Similar conclusions can be made
concerning the amine, 6. Significant amounts of 6 are
detected only at pH e2.0, and it is always a minor
product compared to 5. At 1.0 e pH e 2.0 some of 6 is
generated from hydrolysis of 4 governed by the rate
constant k3, but in more acidic solutions all of 6 appears
to be generated directly from 1.
Figure 3 shows the pH dependence of the product
yields for the four major reaction products 3, 4, 5, and 6.
In this figure yields for 5 and 6 are the yields of those
products obtained directly from 1 without the interme-
diacy of 3 and 4. The yields of 3 and 4 are extrapolated
from the kinetic data in the pH range in which they
exhibit significant decomposition, so 5 and 6 generated
from hydrolysis of 3 and 4 are included as part of the
yields of 3 and 4. Recovery is significantly below 50% at
1 < pH < 5. The yield of one product, 5, varies with pH
in a manner that is directly related to the kinetic results.
If it is assumed that 5 is generated only by the kA process,
ion. The N-sulfonatooxyacetanilides, which also generate
N-acetylnitrenium ions, exhibit ∆Sq of the same sign and
similar magnitude in the same solvent system.9
The kA term shows that 1H+ is (8.0 ( 0.5)-fold more
reactive than 1 toward spontaneous decomposition in our
solvent system. This behavior has not been observed
previously. The conjugate acids of the Glu-P-1 and Glu-
P-2 derivatives 7a and 7b are subject to acid-catalyzed
decomposition, but no spontaneous decomposition of the
protonated species was reported.3 If kA is included in a
fit of the data previously reported for 7a , this term is
not statistically significant at the 95% confidence level.
Such a term is statistically significant at the 95%
confidence level in a fit performed on the previously
reported data for 7b, but it is kinetically insignificant.
Its magnitude is 100-fold smaller than that of the kB term
for 7b, and it is always dominated by the acid-catalyzed
term under conditions in which 7bH+ is present in
significant concentrations. Spontaneous uncatalyzed de-
composition of 7a H+ and 7bH+ is clearly not important
if it occurs at all.
Figure 2 shows the time course for generation of the
hydroxamic acid 3 and the hydroxylamine 5 at pH 1.10.
The majority of 5 (70%) is formed rapidly from 1 without
an apparent intermediate and with a rate constant that
is consistent (Tables S-1 and S-2) with the magnitude of
(9) Novak, M.; Pelecanou, M.; Roy, A. K.; Andronico, A. F.; Plourde,
F. M.; Olefirowicz, T. M.; Curtin, T. J . J . Am. Chem. Soc. 1984, 106,
5623-5631.
(10) Novak, M.; Roy, A. K. J . Org. Chem. 1985, 50, 571-580.
J . Org. Chem, Vol. 68, No. 26, 2003 9877

Novak and Nguyen
TABLE 2. 18O An a lysis of 3 a n d 5 Gen er a ted fr om 1 a t
p H 1.10 in 47.6 m ol % 18O-En r ich ed H2O a n d in Or d in a r y
H2O
SCHEME 1
expt
no.
100% × (M + 3)/
compd
experimenta
[(M + 1) + (M + 3)]b
3
3
3
3
5
5
5
5
1
2
3
4
1
2
5
6
1 in 18O-enriched H2O
1 in ordinary H2O
5.76 ( 0.26
0.88 ( 0.26
2.63 ( 0.40
0.66 ( 0.08
38.71 ( 0.08
1.07 ( 0.23
1.16 ( 0.05
1.33 ( 0.37
3 in 18O-enriched H2O
3 in ordinary H2O
1 in 18O-enriched H2O
1 in ordinary H2O
5 in 18O-enriched H2O
5 in ordinary H2O
a
All compounds were incubated in the pH 1.1 solvent at 40 °C
for 6.5 h before neutralization with 0.02 M H2O4-/HPO42- pH 6.8
buffer, extraction with CH2Cl2, solvent evaporation. Residue was
b
taken up in MeOH immediately prior to analysis. Calculated
from integrated peak intensities for the (M + 1)+ and (M + 3)+
ions obtained by LC/MS analysis, using an APCI source.
its yield should vary with pH according to eq 2
% of 5 ) f510(-x)/(10(-x) + KakB/kA)
(2)
where x ) Ho or pH. A least-squares fit of the percent
yield data for 5 as a function of pH provides f5 ) (90 (
3)% and KakB/kA ) (0.12 ( 0.02) M (Table 1). The value
of KakB/kA derived from the kinetic fit is (0.15 ( 0.03) M.
It appears that 5 directly formed from 1 is generated
exclusively by the kA process.
A mechanism for the formation of 3 and 5 that is
consistent with the experimental data is shown in
Scheme 1. The hydroxamic acid, 3, appears to be gener-
ated from the neutral species 1 because it is not detected
at pH <1. The 18O exchange data are consistent with a
simple acyl-transfer reaction with the hydroxamic acid
serving as the leaving group. This acyl-transfer reaction
has not previously been observed for 7a , 7b, or related
heterocyclic esters.3 These esters are more reactive
toward N-O bond cleavage than 1, so a slow acyl-tranfer
reaction would not be detected for these compounds. Even
in the case of 1, 3 is a minor product with a yield that
does not exceed 15%. The data show that 5 not produced
by hydrolysis of 3 is generated entirely from 1H+. The
pyridyl proton may catalyze an intramolecular nucleo-
philic attack of the ester carboxyl O on the amide that is
initiated by attack of H2O on the ester. This would
generate the cyclic intermediate 8. Reversible N-O bond
cleavage in this species to generate 9 could account for
the 18O exchange with solvent. The exchange in excess
of that predicted from reversible generation of 9 (23.8%
from 47.6 mol % of 18O-enriched H2O) can be accounted
for if 9 reversibly lost H2O to the solvent under the acidic
reaction conditions. Protonation of the N of the 1,4,2-
dioxazolidine ring of 8 would lead to decomposition of 8
generating the hydroxylamine, 5. Other mechanisms
could be written for this process, but they all must be
consistent with spontaneous decomposition of 1H+, 18O
exchange into the hydroxylamine product, and the lack
of involvement of the hydroxamic acid as an intermediate.
This spontaneous reaction of the protonated ester has not
previously been observed with 7a or 7b.3 The reason for
this could be the unique pyridyl protonation site of 1.
Both 7a and 7b are predominately protonated at a
different site (see 7a H+ and 7bH+) that would not allow
for the cyclic transition state shown in Scheme 1.3
The chemical processes leading to 3 and 5 during the
decomposition of 1 were probed by experiments per-
formed in 47.6 mol % of 18O-enriched H2O at pH 1.10 and
40 °C. The 18O content of 3 and 5 was calculated from
the peak intensities for the (M + 3)+ and (M + 1)+ ions
produced from an APCI source of an LC/MS. All experi-
ments were run for 6.5 h, the time at which the maximum
yield of 3 is obtained at this pH. At this reaction time
96% of 5 is generated by the faster process (k1) that does
not involve 3 according to the kinetic data. Experimental
results are summarized in Table 2. Analysis shows an
excess incorporation of (4.88 ( 0.37)% of 18O into 3 formed
from 1 in the 18O-enriched H2O (expt no. 1-expt no. 2 in
Table 2). Control experiments show that (1.97 ( 0.41)%
of this is accounted for by 18O exchange into 3 during
the experiment (expt no. 3 - expt no. 4). The difference,
(2.91 ( 0.55)%, represents the maximum excess 18O
incorporation into 3 during the decomposition of 1. This
is a maximum because some 18O incorporation into the
amide carboxyl O of 1 may occur during the experiment.
This would be transferred to the carboxyl O of 3 and
would contribute to the (M + 3)+ peak intensity. The
small amount of 1 present at the 6.5 h reaction time and
the low intensity of its (M + 1)+ peak precluded direct
measurement of 18O incorporation into 1. In any case,
this result could not distinguish between 18O incorpora-
tion into the amide carboxyl O or the ester carboxyl O.
The results show that a minimum of 94% of 3 is
generated by a process that involves no 18O exchange.
In contrast, considerable 18O exchange occurs into 5.
An excess 18O incorporation of (37.64 ( 0.24)% is observed
(expt no. 1 - expt no. 2). None of this can be attributed
to 18O exchange into 5 after it is generated (expt no. 5 -
expt no. 6). The results show that 79% of the theoretical
maximum exchange occurs during formation of 5.
9878 J . Org. Chem., Vol. 68, No. 26, 2003

N-Acetoxy-N-acetyl-2-amino-R-carboline
-
F IGURE 4. Plots of kobs for the disappearance of 1 vs [N3
]
at pH 3.7 (2) and pH 6.9 (1). The lines are a linear least-
squares fit of the data. The values of kaz′/k1 are the slope/
intercept of the least-squares fit.
F IGURE 5. Plot of the yields of 3 (2), 4 (1), and 10 (insert,
b) as a function of [N3-] at pH 6.9. Data were fit to eqs 3 (for
3), 4 (for 4), and 5 (for 10) as described in the text.
TABLE 3. Azid e Tr a p p in g Resu lts
by this reaction would be e5% if the sensitivity of these
esters to the second-order reaction with N3- is about the
same as that for 1.
a
a
10-2kaz′/k1 (M-1
)
10-2kaz/kc (M-1
)
pH
from kinetics
from 3
from 4
from 10
The amide, 4, disappears with increasing [N3-] to a
greater extent than can be accounted for by the kaz′/k1
partitioning. A new product also appears in the presence
of N3-. This material was not isolated because of its low
yield, but mass spectrometric data taken by LC/MS are
consistent with an N3--adduct represented by structure
10. The observed yields of both 4 and 10 as a function of
[N3-] were fit to eqs 4 and 5, respectively.
6.9
3.7
2.4 ( 0.2
7.2 ( 1.2
3.3 ( 0.7
5.7 ( 0.8
6.0 ( 0.8
2.9 ( 0.7
12 ( 3
a
Ratios reported with their estimated standard deviations from
fits to the equations shown in the text.
Figure 4 shows that kobs varies linearly with [N3-] in
0.02 M H2PO4-/HPO4 buffers (pH 6.9) and in 0.02 M
2-
AcOH/AcO- buffers (pH 3.7). Previously, rate constants
for the decomposition of other esters of heterocyclic
[4] ) [4]max(1/(1 + (kaz′/k1)[N3-]))(1/
(1 + (kaz/kc)[N3-])) (4)
[10] ) [10]max(1/(1 + (kaz′/k1)[N3-]))((kaz/kc)[N3-]/
(1 + (kaz/kc)[N3-])) (5)
hydroxamic acids, including 7a and 7b, have been shown
to be insensitive to [N3-].3 The greater sensitivity to [N3
]
-
at lower pH indicates that 1H+ reacts more readily with
-
N3 than does 1, but the N3--dependence is significant
under both pH conditions. Second-order rate constants
-
for reaction of N3 with 1 and 1H+, kazB′ and kazA′, are
2.4 × 10-2 M-1 s-1 and 2.9 × 102 M-1 s-1, respectively,
In each case the value of kaz′/k1 employed was that
obtained from the fit of the yield of 3 to eq 3. The other
two parameters ([4]max or [10]max and kaz/kc) were adjust-
able least-squares parameters. The fits at pH 6.9 are
shown in Figure 5. Equations 4 and 5 are consistent with
a second N3--dependent partitioning process that occurs
after the original kaz′/k1 partitioning.
assuming the pKa for 1H+ shown in Table 3 and k1 at
both pH values of ca. 1 × 10-4 s-1
.
The nature of this second-order process is revealed by
the product data in Figure 5. The hydroxamic acid, 3, is
produced at the expense of the amide, 4, as [N3
increases. A fit of [3] at pH 6.9 to eq 3, where [3]o, [3]max
and kaz′/k1 are adjustable parameters, is shown in Figure
5. The parameter [3]o is required because there is a small,
-
]
,
In the previous study involving 2, pH-dependent trap-
ping of an intermediate nitrenium ion, 11, and its
quinonoid-like conjugate base, 12, was invoked to explain
[3] ) ([3]o + (kaz′/k1)[N3-][3]max)/(1 + (kaz′/k1)[N3-])
the effects of N3 and Br- on product distributions and
-
(3)
but measurable, yield of 3 in the absence of N3-. The
value of kaz′/k1 obtained from the fit at pH 6.9 of (330 (
70) M-1 is in reasonable agreement with the rate constant
ratio calculated from the kinetic data (Table 3). At pH
-
3.7 similar agreement is obtained. It is clear that N3
reacts selectively with the acetoxy group of 1 or 1H+ to
catalyze a C-O bond cleavage reaction that generates
3. Presumably this reaction has not been observed with
other esters of heterocyclic hydroxamic acids because
they exhibit much more rapid N-O bond cleavage than
does 1 (30- to 1300-fold larger rate constants under the
-
same conditions).3 At the low N3 concentrations used
in these studies (e10 mM), the rate accelerations caused
J . Org. Chem, Vol. 68, No. 26, 2003 9879

Novak and Nguyen
quinonoid-like conjugate base.2,11 The N-acetyl group of
SCHEME 2
app
15 should lower pKa′
below that of 11. Therefore, the
observed value of kaz/kc at pH 6.9 should be a good
approximation to kaz°/kc°. The values of kaz/kc obtained
from 4 and 10 at pH 6.9 (Table 3) are in modest
agreement (within 2 standard deviations). The average
value of 450 M-1 is ca. 2.7-fold smaller than the value of
1.2 × 103 M-1 reported for kaz°/kc° for 12.2
identities.2 The major reaction product of 2 in the absence
-
of N3 was the reduction product 6 (ca. 50% at neutral
pH, and 20% at pH <4.0).2 The analogous product 4 is
the major observed product of 1 from pH 3.0 to 7.0, and,
similar to 6, its yield decreases as pH decreases in that
range. The hydroxylamine ester 2 was not subject to an
N3--dependent second-order rate acceleration, but the
yield of 4 did decrease in the presence of N3-, and the
two N3--adducts 13 and 14 were formed in its place.2 The
-
N3 trapping and results of trapping experiments per-
formed with H- donors suggested that 6 was produced
by reduction of singlet 11 and, predominately, 12, not a
triplet or radical species.2 The species responsible for the
reduction was not identified, but it could be other reaction
products generated by 2.2
The N-acetyl group lowers the pKa of 1H+ by 1.8 units
compared to that for 2H+. If this is taken as an estimate
app
for the effect of the N-acetyl group on pKa′
for 11, the
estimated pKa′ app for 15 is ca. 3.1. If that value of pKa′ app
is substituted into eq 6 with the observed value of kaz/ks
The effects of N3- on the reactions of 1 can be explained
by the mechanism of Scheme 2. The nitrenium ion
component of this mechanism is identical with that
at pH 3.7 (Table 3),and kaz°/kc° estimated above, kaz+/kc
+
is estimated to be 4.2 × 103 M-1. The error limit on this
number is estimated to be (50%, but it is only 8-fold
-
previously proposed for 2.2 Since N3 clearly increases
the rate of decomposition of 1 and 1H+ to generate the
hydroxamic acid, 3, nucleophilic catalysis of acyl transfer
by N3- on both 1 and 1H+ is included. This process is in
competition with the k1 process for 1 that leads predomi-
smaller than kaz+/kc for 11.2 This result suggests that
+
the effect of the N-acetyl group on the selectivity of 15 is
similar to that previously observed for carbocyclic N-
acetyl-N-arylnitrenium ions.6,7
-
nately, but not exclusively, to N-O bond cleavage. N3
Other products attributed to the nitrenium ion path-
way for 2 included the dimers 18a -c that appeared to
be generated by attack of the amine, 6, on the nitrenium
ion 11 or its conjugate base 12.2 These dimers are
unusual products that are not observed in other studies
involving heterocyclic or carbocyclic N-arylnitrenium ions
because significant concentrations of the amine or amide
reduction products are usually not present during these
reactions.1,3,6,7 The dimers appear to be formed because
of the propensity for reduction of 11 and 12, and their
relatively long lifetime, so that they can be nucleophili-
cally trapped by their own reduction product.2 It has been
shown in other studies that aromatic amines such as
also affects distribution of products derived from the
nitrenium ion, 15, and its conjugate base, 16. The
conjugate acid-base pair is invoked because there is
evidence, based on data taken at two pH values (Table
3), that the extent of N3--trapping at this stage is pH-
dependent. It is not possible to determine the two limiting
+
trapping ratios, kaz+/kc and kaz°/kc°, with only two data
points without knowledge of -log(Ka′kc°/kc+), pKa′
,
app
obtained from a fit of kaz/kc to [H+] (eq 6 ).2 For 11, pKa′ app
kaz/kc ) (kaz+[H+]/kc+ + kaz°Ka′ app/kc°)/([H+] +
Ka′ app) (6)
(11) McClelland, R. A.; Licence, V. E. Arkivoc 2001, 12, 72-81
(http://www.arkat-usa.org/ark/journal/Volume2/Part3/Tee/OT-286C/
OT-286C.pdf).
is 4.9, while it is 7.7 for the related 2-carbazolylnitrenium
ion, 17, that also undergoes deprotonation to a neutral
9880 J . Org. Chem., Vol. 68, No. 26, 2003

N-Acetoxy-N-acetyl-2-amino-R-carboline
products were examined. For the UV method, absorbance at
305 or 275 nm was plotted against time. The UV and HPLC
data were fit by nonlinear least-squares methods to the
standard first-order rate equation or to consecutive first-order
rate equations. The HPLC conditions were the following: UV
detection at 320 nm, 55/45 MeOH/H2O eluent buffered with
0.025 M 1:1 HOAc:KOAc, C-18 reverse-phase column, 1 mL/
min flow rate.
Similar procedures were followed for kinetic measurements
involving 3 and 4. To carry out the reaction, 15 µL of a 0.002
M stock solution of 3 or 4 in DMF was injected into 3 mL of
the appropriate solution incubated in a 40 °C water bath to
obtain an initial concentration of 1 × 10-5 M. HPLC data were
fit by nonlinear least-squares methods to the standard first-
order rate equation. Rate constants for very slow reactions (kobs
< 10-6 s-1) were determined by initial rates methods from the
slope of a plot of reactant HPLC peak area vs time for the
first 5% of the reaction.
The products of the decomposition of 1 were monitored by
HPLC and LC/MS under conditions identical with those used
for the kinetic studies. The major products 3, 4, 5, and 6 were
identified by co-injection with authentic compounds, and by
comparison of mass spectra obtained by LC/MS with those of
the authentic compounds. When 1 decomposed at pH 3.0 and
5.0, a small amount of a dimer (19) was observed by HPLC
with 75/25 MeOH/H2O as eluent. LC-MS (APCI, positive ion
mode) C26H21N6O2 (M + H)+ calcd m/e 449.15, found 449.20.
At pH 6.9, during the decomposition of 1 in the presence of
N3-, the azide adduct 10 was observed. LC-MS (ESI, positive
ion mode) C13H11N4O (M - N2 + H)+ calcd m/e 239.28, found
239.70.
aniline and N,N-dimethylaniline are excellent traps for
N-arylnitrenium ions.12 There is another recent example
of a long-lived cation (R-(N,N-dimethylthiocarbamoyl)-
4-methoxybenzyl cation) that reacts with its own elimi-
nation product to form a dimer because of the long
lifetime of the cation.13 Increasing solvent nucleophilicity
suppresses the formation of this dimer.13 In the present
study a product with HPLC retention time and mass
spectrum consistent with the dimeric structure 19 was
detected by LC/MS in reaction mixtures of 1 in the pH
range 3.0 to 5.0. This provides further evidence that the
nitrenium ion path occurs for 1 despite the many differ-
ences in the chemistry of 1 and 2.
In conclusion, the addition of the N-acetyl group to
form 1 results in significant suppression of the rate of
the N-O bond cleavage process in H2O so that competing
reactions that are not observed in other model esters for
carcinogens derived from heterocyclic amines can be
detected. Included among these are a slow uncatalyzed
acyl-transfer to the solvent, a more rapid acyl-transfer
catalyzed by N3-, and an apparently unique spontaneous
decomposition of the conjugate acid of 1, 1H+, that may
be facilitated by an intramolecular proton transfer.
Although the nitrenium ion path is kinetically suppressed
by the electron-withdrawing N-acetyl group, it still
appears to be the dominant process at neutral pH in the
absence of strong nucleophiles such as N3-. The proper-
ties of the nitrenium ion 15 and its conjugate base 16
appear to be similar to those of the deacetylated ana-
logues 11 and 12.
18O Exch a n ge in to 3 a n d 5. The 18O-enriched solution (1
mL) was made with 0.10 mL of 1 M HClO4, 0.05 mL of CH3-
CN, 0.375 mL of distilled H2O, and 0.475 mL of 95 atom % of
18O-enriched H218O. An ionic strength of 0.5 was maintained
with 0.0562 g of NaClO4. Reactions were performed as
described above for the kinetics experiments. After 6.5 h, the
reaction solution was neutralized to pH 5.0 by addition of 1.5
mL of 0.02 M phosphate buffer (pH 6.80) and then extracted
with 3 × 1 mL of CH2Cl2. The CH2Cl2 was evaporated, without
predrying. The solid was immediately redissolved in MeOH
and subjected to LC/MS to analyze the 18O content of 3 and 5.
Controls in which 3 and 5 were incubated in the 18O-enriched
solution for 6.5 h were performed, as were controls for 1, 3,
and 5 in ordinary unenriched pH 1.1 solution. In each of these
experiments, analyses for the 18O content of 3 and 5 was
performed by obtaining the integrated peak intensities for the
(M + 1)+ and (M + 3)+ ions. For 3, LC/MS (APCI, positive ion
mode) C13H12N316O2 (M + H)+ calcd m/e 242.26, found 242.10,
and C13H12N318O16O (M + H)+ calcd m/e 244.27, found 244.00.
For 5, LC/MS (APCI positive ion mode) C11H10N316O (M + H)+
calcd m/e 200.08, found 200.10, and C11H10N318O (M + H)+
calcd m/e 202.09, found 202.10.
Exp er im en ta l Section
Gen er a l P r oced u r es. The synthesis and characterization
of 1, 3, and 4 are described in the Supporting Information.
All other isolated compounds are described in the literature.14
All salts used for preparation of buffers were reagent grade.
The purification of CH3CN and DMF was carried out according
to the method described elsewhere.15 Water used in all the
experiments and kinetic studies was distilled, deionized, and
distilled again. All other reagents and solvents were reagent
grade and distilled. General procedures for preparing solutions,
measuring rate constants by UV and HPLC methods, and ana-
lyzing products by HPLC methods have been described else-
where.2,6,9,10
Kin etics a n d P r od u ct Stu d ies. The kinetic study for 1
was performed in 5 vol % of CH3CN-H2O solutions, µ ) 0.5
(NaClO4), T ) 40 °C, unless otherwise stated. To maintain pH,
NaH2PO4/Na2HPO4 (pH 5.5-7.5), AcOH/AcONa (pH 4.0-5.0),
and HCOOH/HCOONa (pH 3.0-3.5) buffers were employed.
HClO4 solutions were used at pH <3.0, and in the Ho region
where ionic strength was not maintained.
To carry out the reaction, 15 µL of a 0.001 or 0.01 M stock
solution of 1 in DMF was injected into 3 mL of the appropriate
solution to obtain initial concentrations of 1 of 5 × 10-6 or 5 ×
10-5 M. The solution was incubated for at least 15 min prior
to initiation of the reaction at 40 °C in either a thermostated
cell (when using the UV spectrophotometer) or a water bath
(when monitoring the reaction by HPLC). For the HPLC
method, peak area vs time data for the starting material and
Ack n ow led gm en t. This work was supported by a
grant from the American Cancer Society (RPG-96-078-
03-CNE). NMR spectra were obtained on equipment
made available through an NSF grant (CHE-9012532)
and upgraded through an Ohio Board of Regents
Investment Fund grant. LC/MS were obtained on
equipment made available through an Ohio Board of
Regents Investment Fund grant. High-resolution mass
spectra were obtained at the Ohio State University
Chemical Instrumentation Center.
(12) Rangappa, K. S.; Novak, M. J . Org. Chem. 1992, 57, 1285-
1290. Novak, M.; Rangappa, K. S.; Manitsas, R. K. J . Org. Chem. 1993,
58, 7813-7821.
Su p p or tin g In for m a tion Ava ila ble: Synthesis and char-
acterization of 1, 3, and 4, and tables of rate constants (Tables
S-1 through S-3) for the decomposition of 1, formation of 3-6
during the decomposition of 1, and the hydrolysis of authentic
3 and 4. This material is available free of charge via the
Internet at http://pubs.acs.org.
(13) Williams, K. B.; Richard, J . P. J . Phys. Org. Chem. 1998, 11,
701-706.
(14) Hibino, S.; Sugino, E.; Kuwada, T.; Ogura, N.; Shintani, Y.;
Satoh, K. Chem. Pharm. Bull. 1991, 39, 79-80. Kazerani, S.; Novak,
M. J . Org. Chem. 1998, 63, 895-897.
(15) Novak, M.; Brodeur, B. A. J . Org. Chem. 1984, 49, 1142-1144.
J O034505U
J . Org. Chem, Vol. 68, No. 26, 2003 9881